The aim of the Medical Knowledge Discovery Research project is to make worldwide health and medical information accessible and useful in the daily lives of patients, physicians, clinicians, researchers, students and health care professionals.

“The effective and efficient application of known worldwide health and medical information and knowledge will have a bigger impact on health and disease than any drug or technology likely to be introduced in the next decade” according to Sir Muir Gray, Director of Oxford University Institute of Health Sciences.

To address the challenges of this ambitious project, the Medical Knowledge Discovery Research Consortium has been set up by the Medical Knowledge Discovery Project with Northwestern University Center for Genetic Medicine, Ontology Works Inc., Science Commons, and University of Maryland Institute of Genome Sciences. The Medical Knowledge Discovery project addresses the following areas with innovative approaches to web-enabled health and medical information and knowledge.

Portal to web based medical knowledge
Medical Knowledge Discovery provides portals to web based health and medical knowledge that facilitate the effective and efficient application of known worldwide health and medical information and knowledge. “The effective application of knowledge will have a bigger impact on health and disease than any drug or technology likely to be introduced in the next decade” according to Sir Muir Gray, Director of Oxford University Institute of Health Sciences.

Associating Medical Content and Context
Associating medical content such as diseases, disorders, and medications with the patient and medical context is critical to achieve this effective application of medical knowledge. Extensive research has been carried out into the association of medical context with medical content and how it is most effectively used to identify the most appropriate web based health and medical knowledge.
A methodology has been defined to map Medical Context Descriptors to Open Biomedical Ontologies. This allows the definition and mapping of Disease Context Descriptors to disease classes so that Medical Knowledge Applications and searches can be developed that vary according to the disease, disorder, or medication on which the user is seeking information and knowledge.

Identifying and ranking sources of web-based medical information
Analysis of a number of disease searches show that approximately 40% of the results come from .com web sites, 26% from medical associations and organizations, 10% from government, 5% from medical libraries and the NIH, and 2% from universities. These web sites are grouped into Source Categories.
Research showed referenced web sites vary per disease class as web sites are often focused on a disease class. From this research it follows that web sites must be ranked and qualified per disease class through a “Medical Information IQ” page ranking to give meaningful results. Web sites are ranked by the amount of information they contain on a particular disease, disease class, and medication. This ranking is measured per disease class through the web site "Medical Information Density" and is expressed as the web site "Medical Information IQ" for a disease class and disease.

Medical Knowledge Applications
The Medical Knowledge Base, mapped from Open Biomedical Ontologies, OBO, and in particular the Disease Ontology, DO, supports a number of Medical Knowledge Applications that associate disease attributes, in plain text, with diseases and disorders. Symptom Search associates plain text primary and secondary symptoms, as described by the user, with likely matching diseases. Diagnostic test results, laboratory tests and ECG results are described in plain text by the user for the Diagnostic Test Search. Genetics search associates genes, proteins, and genetic markers with diseases.

Clinical Knowledge Applications
The Clinical Knowledge Portal is designed for clinicians, physicians, researchers, and students to consult key clinical knowledge and information on patient care and clinical research. Answering Clinical Questions research showed that 20% of resulting documents from a clinical question search have some relevance to the clinical question and 10% are extremely relevant. Results from Clinical Knowledge Applications and Clinical Question Search are ranked by Clinical Relevance Score. This score is calculated from the frequency and proximity of clinical question keywords in the resulting clinical documents and web pages.

Viewing Medical Ontologies
The Open Biomedical Ontologies and the Disease Ontology provide the relationships between diseases and a comprehensive view and navigation of diseases which is unsuited for user references. Ontology Views allow user groups to define views of the comprehensive ontology and relationships that are suited to their needs. Different Ontology Views can be created for healthcare consumers, patients, physicians, specialists, healthcare professionals, academics, and students. Ontology Views can also be created for constituencies within these groups who require specialized Ontology Views on a subset of the ontology.